Modern agriculture increasingly relies on data-driven systems to address challenges such as climate uncertainty, inefficient resource usage, and variability in crop productivity. This study proposes a hybrid Digital Twin framework that integrates sensor-based environmental monitoring, process-oriented crop growth modeling, and machine learning techniques to enable intelligent agricultural decision-making. The framework combines key environmental variables, including temperature, rainfall, humidity, soil moisture, and solar radiation, with crop growth indicators such as leaf area index and biomass to simulate crop behavior and predict yield outcomes. To enhance model robustness under limited data availability, a physics-guided synthetic data generation approach is incorporated. In addition, a feedback-driven updating mechanism continuously refines model parameters based on prediction discrepancies, improving system adaptability over time. Experimental evaluation demonstrates that the proposed hybrid approach enhances predictive accuracy and supports efficient resource management. The results highlight the potential of Digital Twin technology in developing scalable, adaptive, and sustainable smart agriculture systems.

Digital Twin, Smart Agriculture, Machine Learning, Crop Yield Prediction, Internet of Things (IoT), Precision Agriculture, Feedback Mechanism, Data-Driven Farming